Re: Possible netfilter-related memory corruption in 2.6.37

[Eric Dumazet <eric.dumazet@gmail.com>]

Le vendredi 18 février 2011 à 19:37 +0100, Patrick McHardy a écrit :
> Am 14.02.2011 17:52, schrieb Patrick McHardy:
> > Am 14.02.2011 17:48, schrieb Eric Dumazet:
> >> I am not sure, but I guess nf_reinject() needs a fix too ;)
> > 
> > I agree. That one looks uglier though, I guess we'll have to
> > iterate through all hooks to note the previous one.
> 
> How about this? Unfortunately I don't think we can avoid
> iterating through all hooks without violating RCU rules.
> 
> 

       /* Continue traversal iff userspace said ok... */
        if (verdict == NF_REPEAT) {
-               elem = elem->prev;
-               verdict = NF_ACCEPT;
+               prev = NULL;
+               list_for_each_entry_rcu(i,
&nf_hooks[entry->pf][entry->hook],
+                                       list) {
+                       if (&i->list == elem)
+                               break;
+                       prev = i;

	
Hmm... what happens if "elem" was the first elem in list ?

We exit with prev = NULL  --> NF_DROP ?

I must miss something...

+               }
+
+               if (prev == NULL ||
+                   &i->list == &nf_hooks[entry->pf][entry->hook])
+                       verdict = NF_DROP;
+               else {
+                       elem = &prev->list;
+                       verdict = NF_ACCEPT;
+               }
        }



--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html